I just got a 90-teraflop (AMD Ryzen AI 7 with NPU and Nvidia 5070 RTX GPU) laptop for $1200, and that made me reflect on the history of computing. This is like having a supercomputer worth tens of millions of dollars and consuming ~10 MW of power in 2005, right on my kitchen desk and completely dedicated to me, 24/7.
The exponential increase in computing has been a constant over the last seven decades, and it is only accelerating with the recent emphasis on AI infrastructure. All of us have benefited in diverse ways from this rapid advancement. Below is my attempt to summarize the progress in hardware and the dramatic reduction in cost. I have superimposed important milestones from my own career where I was able to harness that wave. This has transformed my work in many positive ways, and the trend is only going to accelerate, requiring everyone to think computationally to navigate this never-ending current. In the interactive chart below, you can add a milestone of significance to you and see where it fits within this exponential growth.
Here is some more background on my milestones:
To contextualize this history, we can plot specific systems from a professional’s career onto this broader performance curve. This exercise transforms an abstract historical trend into a tangible personal journey through the evolution of computing.
1992: The First Rung on the Ladder
- The Machine: Intel 486 at IIT Kharagpur. Access to a 486-based system in 1992 represented the high end of personal computing. Such a system would have likely featured an Intel 486DX-33 or DX-50 processor. Based on benchmark data, its computational power can be estimated at approximately 4 to 6 MFLOPS. This placed it firmly in the “MegaFLOPS” era for desktop machines—a level of power that was revolutionary for an individual user, even as it was dwarfed by the GigaFLOPS-scale supercomputers of the day.
- The Experience: This was a shared system where I first coded a tridiagonal solver and presented a paper at the Tryst conference at IIT Delhi in 1992. I got 3rd place for my talk and I remember getting a cash prize. I was so proud to buy my mom a bag with my first earned paycheck.
1994: Diving into High-Performance Computing
- The Machine: KSR1 at Georgia Tech. The Kendall Square Research KSR1 was a highly innovative MPP supercomputer built from proprietary 64-bit RISC CPUs. Each processor ran at 20 MHz with a peak double-precision performance of 40 MFLOPS. A common 32-processor configuration would thus have a theoretical peak performance of 32×40 MFLOPS=1.28 GFLOPS. This machine occupied a powerful middle ground, far exceeding any desktop capability but not quite reaching the top tier of the largest government systems.
- The Experience: This is where I started my journey in High-Performance Computing (HPC), modeling underwater bubble explosions, LES of single and two-phase combustion, and more.
1995-1999: Riding the Teraflop Wave
- The Machines: Fastest US DoD Machines. This period saw an inflection point where the dominance of U.S. Department of Defense (DoD) systems was being challenged by Department of Energy (DOE) laboratories, especially through the ASCI program and the race to the teraflop. Through the DoD HPC Modernization Program (HPCMP), I had a grand challenge project on modeling two-phase reacting flows and had access to the fastest DoD machines of that time. Not only did I develop my own code, but I also learned to parallelize on distributed and shared memory systems.
1999-2015: At the Forefront with ORNL
- The Machines: Fastest Computers at ORNL. Oak Ridge National Laboratory has consistently hosted world-leading computational resources. I had the privilege of working with leading applied math and computer science researchers to scale various software, including MFIX, to model some of the most complex phenomena. The large datasets generated on these supercomputers also led me to think about big data and machine learning to construct reduced-order models, even before those topics became buzzwords.
- The period began with the IBM “Eagle” system (1999-2000), the first teraflop-class system dedicated to open science, with a peak performance of approximately 1.5 TFLOPS.
- This was followed by the Cray “Jaguar” system. After several upgrades, it was ranked #1 on the TOP500 list in November 2009 with a theoretical peak of 2.3 PetaFLOPS.
- In 2012, ORNL unveiled “Titan,” a Cray XK7 system that pioneered the large-scale hybrid CPU/GPU architecture, debuting at #1 on the TOP500 list with a staggering theoretical peak performance of 27 PetaFLOPS.
Below is a screenshot of the interactive webpage, and here is the link for you to try it out: https://ekta.net/LI/exponential-compute.html
